Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
An international team of researchers has created a futuristic tool to examine footprints left by dinosaurs in our ancient past. The AI-powered Dinotracker app can identify dinosaur footprints in moments.
This research comes from a joint project between the Helmholtz-Zentrum Research Center in Berlin and the University of Edinburgh in Scotland. Published by Proceedings of the National Academy of Sciences The paper on monday.
Don’t miss any of our unbiased technical content and lab reviews. Add CNET As Google’s preferred source.
Identifying dinosaur species by footprint is not always easy. The footprint is hundreds of millions of years old, and is often preserved in layers of rock that have moved over the ages since the track was laid.
Also, we still have a lot to learn about dinosaurs, and it’s not always clear which species left a mark. Subjectivity or bias can play a role when determining them, and scientists do not always agree with the results.
Gregor Hartmann of Helmholtz-Zentrum, who led the project, told CNET that the research team sought to remove this tendency from the identification process by developing an algorithm that could be unbiased.
“We bring a mathematical, unbiased point of view to the table to help human experts interpret the data,” Hartmann said.
The researchers trained the algorithm on thousands of real fossil footprints, as well as millions of simulated copies that can recreate “natural deformations such as pressure and shifting edges.”
How artificial intelligence is used on dinosaur tracks
The system was trained to focus on eight key characteristics of dinosaur footprints, including the width of the toes, the position of the heel, the surface area of the foot in contact with the ground, and the distribution of weight across the foot.
The AI tool uses these features to compare new footprints to existing fossils, and then determine which dinosaur was most likely responsible for the print.
The team tested it against the ratings of human experts, and found that the AI agreed with them 90% of the time.
Hartmann explained that the AI system is “unsupervised.”
“We don’t use any labels (such as birds, theropods, ornithopods) during training. The network has no idea about that,” Hartmann said. “Only after training do we compare how the network encodes the silhouettes and compare them to human labels.”
Hartman said the hope is that Dinotracker will be used by paleontologists and that the AI tool’s data set grows as it is used by more experts.
Birds vs dinosaurs
Using Dinotracker, researchers have already uncovered some interesting possibilities about bird evolution. Analyzing footprints more than 200 million years old, the AI found strong similarities to the foot structures of extinct and modern birds.
One possibility, the team says, is that birds arose tens of millions of years earlier than we thought. But it’s also possible that early dinosaur feet looked remarkably like bird feet.
Hartmann points out that this evidence is not enough to rethink bird evolution, because the skeleton is the “real evidence” of early birds.
“It is necessary to take into account that over millions of years, many different things can happen to these tracks, from the moisture level of the clay where it was created, over the substrate on which it was created, to subsequent erosion,” he said. “All of this could significantly change the shape of the fossilized track we find, and ultimately make the footprints very difficult to interpret, which was the motivation behind our study.”
Dinotracker is Available for free On github. It’s not a download-and-use format, so you’ll have to know a little about the software to run it.